Journal of Materials Processing Technology 209 (2009) 42924311

Contents lists available at ScienceDirect

Journal of Materials Processing Technology

journal homepage: www.elsevier.com/locate/jmatprotec

Internal ductile failure mechanisms in steel cold heading process

A. Sabih a, , J.A. Nemes b
a
Department of Mechanical Engineering, McGill University, 817 Sherbrooke Ave West, Montreal, Quebec, Canada H3A 2K6
b
Engineering Division, Pennsylvania State University-Great Valley, 30 East Swedesford Road, Malvern, PA 19355-1443, United States

a r t i c l e i n f o a b s t r a c t

Article history: The occurrence of internal ductile failure in cold-headed products presents a major obstacle in the fast
Received 22 February 2008 expanding cold heading (CH) industry. This internal failure may lead to catastrophic brittle fracture under
Received in revised form 5 November 2008 tensile loads despite the ductile nature of the material. Comprehensive testing and investigation method-
Accepted 7 November 2008
ologies were used to this work to reveal the complicated interplay of process and material parameters
contributing in the initiation and propagation of internal ductile failure in six CH quality AISI steel grades.
Keywords:
The metallurgical and microscopic investigations showed that internal ductile failure occurs pro-
Ductile failure
gressively by void nucleation and growth mechanisms with increasing plastic strain inside the highly
Adiabatic shear band
Cold heading
localized adiabatic shear bands (ASBs). The void nucleation occurs by decohesion at second-phase par-
Void nucleation ticles, inclusionmatrix interfaces, grain boundaries and by particle or inclusion cracking. Therefore,
Inclusions the number and morphology of any inclusions and second-phase particles are key factors in material
formability.
The metallurgical investigations showed that under compressive loading conditions, the nature of the
metal ow pattern promotes different rates of material ow around the inclusions and stringers which
supports decohesion and void nucleation since the early stages of deformation. At advanced stages of
deformation, the metal ow pattern contributes to the ASB localization in supporting void growth and
coalescence along the band leading to narrow void sheets.
All tested materials in this work experienced ductile failure by void nucleation and coalescence, forming
cracks along the ASBs. The ductile failure of each material was the result of the contribution of all the
mechanisms of void nucleation at the inclusionmatrix interface, second phasematrix interface and at
the grain boundaries. However, the level of contribution of each mechanism in the nal ductile failure
varied depending on material properties and their microstructure.
2008 Elsevier B.V. All rights reserved.

1. Introduction reported that the decision to use the CH process to manufacture

The occurrence of any failure is a major limitation governing the
limits of any forming process. Therefore, understanding the fail-
ure mechanisms and the complex interplay of process and material
parameters in the failure occurrence in metal forming operations
has attracted the attention of many researchers for more than six
decades. The knowledge of different failure mechanisms is crucial
to improve product quality and design methodologies.
Barrett (1997) considered in his report about the fasteners that
the cold heading (CH) process is one of the most important multi-
stage metal forming processes because of its many advantages over
machining of the same part, including high productivity for com-
plex nal shapes, minimum material waste and increased tensile
strength from cold working. Chitkara and Bhutta (2001) who stud-
ied the near-net shape heading of splines and solid spur gear forms
Corresponding author. Tel.: +1 514 398 8546; fax: +1 514 398 7365.
E-mail address: amar.sabih@mcgill.ca (A. Sabih).
certain products depends on the complexity in the shape of the
head and the material employed and the possibility of internal and
external failure occurrence.
Currently, the CH industry favors using faster headers, reducing
the number of manufacturing stages, and producing high-strength
fasteners without nal heat treatment. To achieve these goals, mod-
ied process designs are required which result in higher strains and
strain rates, which cause two familiar types of ductile defects in the
cold-headed products. The rst is the external oblique or longitu-
dinal crack caused by the exhaustion of the material ductility. The
second failure was reported by Bai and Dodd (1992) in their study
for the adiabatic shear band (ASB) phenomenon. They reported that
the ASBs may lead to internal crack. Okamoto et al. (1973) stated in
their study of different forming processes that this type of failure
may result in splitting of the fasteners heads as shown in Figs. 13,
respectively.
Many researchers have focused on surface defects. Cockroft and
Latham (1968) and Lee and Kuhn (1973) studied this defect in upset-
ting and presented different criteria to predict it. Nickoletopoulos

0924-0136/$ see front matter 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmatprotec.2008.11.023
 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311 4293

(2000) studied the surface defect in CH process and reached to a failure extensively, it can be visually inspected and accounted for
conclusion that CockroftLatham criterion can be used to predict in the process design of the component for removal by means of
the surface defect in CH process with a reasonable accuracy. In machining or trimming. Thus far, internal ductile failure due to the
addition to the large number of studies that covered this type of ASB phenomenon in CH process has not received the same atten-

Fig. 1. Cold-headed industrial bolt: (a) without defect, using material 1, (b) shank of the defective bolt (material 2), (c) separated industrial bolt head due to an internal crack
(material 2), and (d) magnied detail of the ductile crack of the separated industrial bolt head (material 2).

Fig. 2. Sectioned bolts revealing material ow inside the bolt head after etching with Frys reagent (cupric chloride 36 g; 145 ml hydrochloric acid; 80 ml water): (a) without
defect bolt (material 1), (b) with defect bolt (material 2).
4294 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311
tion. The clear need for an in-depth understanding of the internal
ductile failure mechanisms which limits the development of new
CH designs and the improvement of cold-headed product quality
motivated this work.
To have a full understanding of the complicated parameters and
mechanisms contributing in the initiation and propagation of the
internal ductile failure, the rst part of this paper will provide a
detailed introduction covering important details about the CH pro-
cess and the ASB phenomenon followed by examples of ductile
failure in cold-headed parts. The second part of this paper will focus
on the experimental and metallographic procedures used to inves-
tigate the failure initiation and propagation. The last part of this
paper presents the detailed result and discussion.
1.1. Cold heading process
The term heading refers to cold, warm, or hot working of metal
employing a static or dynamic impact compressive force causing the
metal at one end of the rod or blank to form a head. Yoo et al. (1997)
dened the cold heading (CH) process as a forming process that is
performed without an external heat source and it involves applying
a force to the free end of a metal workpiece contained between a
die and a punch, by one or several blows of the punch. The CH pro-
cess is a high-rate deformation process with strain rates exceeding
100 s1 . Kawashima (1992) reported that the CH process is used
to produce a large variety of components such as fasteners, studs,
small shafts, and other machine parts. The automotive, construc-
tion, aerospace, railway, mining and electrical product industries
are major consumers of such parts.
Matsunaga and Shiwaku (1980) concluded in their study of man-
ufacturing CH quality wire rods and wires that material properties
and process parameters signicantly affect the heading results.
Moreover, the CH material must meet two sets of requirements.
The rst set concerns the materials ability to be formed with-
out failure while the second set is linked to the properties of the
nal product. Sarruf et al. (1999) stated that in order to form
without failure, CH wire should be free from surface defects and
internal inclusions. Nickoletopoulos et al. (2001) reported that the
chemical composition of CH wires plays a signicant role in the
material ductility and the nal properties of the products. Low- and
medium-carbon steels (0.060.30% C), are commonly used in the
CH process, as they offer good ductility. Also, Nickoletopoulos et al.
(2001) referred that the CH material microstructure is usually com-
prised of a ferrite matrix with varying amounts of lamellar pearlite.
Fig. 3. Cold-headed brass bolt experienced ductile failure in service along the ASB: (a) and (b) sectioned failed bolt, (c) crack surface in a failed bolt (Reitz, 2005) Reprinted
with permission of ASM International (www.asminternational.org).
A spheroidized microstructure provides the desired ductility for
CH of complex geometries. The required nal strength properties
are usually achieved through alloy addition and/or heat treatment
(quenching and tempering).
1.2. ASB phenomenon
Wright (2002) dened the ASB as a two-dimensional, narrow,
nearly planar region of very large shearing that occurs in metals and
alloys experiencing intense dynamic loading (as in the CH process).
When the band is fully developed, the two sides of the region are
displaced relative to each other. However, the material still retains
full physical continuity from one side to the other. The thickness
of the most heavily sheared region might be few tens of microns
or less, and its length might extend many millimeters or centime-
ters. Generally, ASBs form under impact loading at high-strain rates
(higher that 102 s1 ) and high strains. With increasing plastic defor-
mation, the work hardening (from increases in strain and strain
rate) results in an increase in the ow stress. However, most of
the plastic work (9095%) is converted into heat causing a local
temperature increase and a ow stress decrease due to thermal or
work softening. Thus, a competing mechanism between the work
hardening and the thermal or work softening commences and con-
tinues in the deformation zone. As the work hardening mechanism
dominates over the work softening at the beginning, an increase in
the ow stress occurs and with continuing deformation, the ther-
mal or work softening mechanism can progressively dominate over
work hardening increases, which then triggers unstable deforma-
tion. Wright (2002) also reported that this instability condition
will force the deformation to localize into a narrower band that
through further localization can lead to nal failure. Cowie et al.
(1989) showed in their study of microvoid formation during shear
deformation of ultrahigh-strength steels that this failure initiates
and develops by progressive nucleation, growth and coalescence
of microvoids and microcracks. A sharp drop in the load-carrying
capability for the deformed zone will associate with microvoid for-
mation causing what is known as the microvoid softening. Once the
failure starts to develop, the combined work softening effect of the
thermal and microvoid softening mechanisms will compete with
the work hardening effect within the band.
Zukas (1990) reported that there are two types of ASBs, namely
the deformed adiabatic shear bands (DASBs) and the transformed
adiabatic shear bands (TASBs). DASBs occur in materials that do
not undergo phase transformation, or when the local temperature
 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311
within the band is not high enough to cause phase transformation.
The damage and fracture process in DASBs involves a number of
metallurgical events involving different steps between void nucle-
ation and crack propagation. The initial phase of damage coincides
with void nucleation at inclusion edges or at grain boundaries due
to the interaction of local stresses and dislocations phenomena. The
damage in the material is then driven by the accumulated plastic
strain and affected by the stress triaxiality. Hambli (2001) reached
an important conclusion that deformation under compressive (neg-
ative) stress triaxiality increases material formability in comparison
with deformation under tensile loading state, i.e. positive stress tri-
axiality. Zukas (1990) declared that the competition between the
stress triaxiality levels reached in the specimen and the high-plastic
strains will determine the failure site.
Daridon et al. (2004) reported that the localization of plastic
deformation in narrow shear bands is a major damage mechanism
that leads to ductile failure during high-strain rate deformation. Lee
et al. (2004) reported that internal cracks occur within the shear
Fig. 4. As-received testing materials microstructure: (a) 1008 steel, (b) 1018 steel, (c) 1038 steel, (d) 1541 steel, (e) 8640 steel, and (f) DP steel.
4295
band, and become the sites of eventual failure. Venugopal et al.
(1997) concluded that this can take place when the material is sub-
sequently subjected to dynamic impact loading during forming or
in service.
TASBs occur in materials that undergo phase transformations,
and are often found in high-strength steel in locations where the
critical temperature for the transformation of ferrite to austenite
(Ac3 ) is exceeded. After high-strain rate forming, the rapid heat dis-
sipation from the ASB to the surrounding matrix material results
in a fast decrease in temperature below the Ms limit. This sud-
den change in temperature leads to austenite transformation into
martensite, yielding the TASBs. TASBs may fracture in a brittle man-
ner in directions normal to the local tensile stress. If the matrix is
sufciently ductile, brittle fracture will be limited to the band, but
sometimes the fracture may extend to the matrix. Rogers (1983)
stated that even if the TASB does not fracture, it remains a brit-
tle fracture path in the middle of the ductile matrix and might
lead to catastrophic failure. Moreover, Klepaczko et al. (1988) have
4296 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311

indicated that shear bands either with or without phase transfor- Table 1
Chemical analysis of the testing materials (wt%).
mations, are considered as a site of fracture initiation.
AISI steel grade C Mn P S Si Cu Ni Cr

1.3. Examples of internal ductile failure in cold-headed products 1008 0.09 0.4 0.018 0.012 0.01 Nil 0.01 0.02
1018 0.16 0.69 0.006 0.006 0.22 0.04 0.08 0.08
1038 0.38 0.82 0.007 0.002 0.19 0.04 0.08 0.08
The following paragraphs present cold-headed products failed 1541 0.37 1.41 0.011 0.018 0.22 0.16 0.08 0.09
by internal ductile fracture along the ASBs. Fig. 1 shows in detail 8640 0.40 0.91 0.009 0.005 0.25 0.14 0.42 0.43
industrial bolts were made of two different 1018 steels from two DP 0.088 1.69 0.009 0.011 0.62 0.04 0.07 0.07
different CH steel providers (material 1 and material 2) using the
same CH sequence. The industrial bolt made of material 1 (Fig. 1(a)) AISI steel grade Mo Sn Al N B V Co

did not exhibit internal cracks while the second bolt made of mate- 1008 0.002 0.002 0.038 0.0046 0.004
rial 2 failed by a crack resulting from the head splitting when a 1018 0.001 0.002 0.002 0.0026 0.0001 0.005
1038 0.001 0.003 0.003 0.0044 0.0001 0.005
low-tensile load was applied. Fig. 1(b)(d) shows the details of the
1541 0.021 0.010 0.004 0.0058 0.0000 0.002
crack inside the head of the bolt made of material 2. The met- 8640 0.217 0.010 0.005 0.0063 0.0000 0.032
allographic study of the sectioned bolts made of the material 2 DP 0.458 0.005 0.003 0.0035 0.007 0.033
(Fig. 2(b)) conrmed the nucleation, growth and coalescence of
microvoids leading to microcracks.
Reitz (2005) investigated the failure of brass bolts after 12 years of 0.002 s1 , the force and the axial displacement values were
in service. The microscopic investigation revealed that the bolts recorded. The engineering stressengineering strain curves were
failed by macrocracks along the highly deformed band inside the constructed from the loadelongation measurements. These curves
bolts head (Fig. 3). were used to calculate the true stresstrue strain curves (Fig. 5) for
In general, the ductile crack occurred inside a highly localized all testing materials.
region of the ASB. The occurrence of ductile failure in material 1
using the same CH multistage design in the part in Fig. 1 raises 2.2. Testing material cleanliness
important questions about the ASB phenomenon and how it leads to
such a catastrophic failure. The ductile failure that took place inside Cleanliness is a measure of the inclusion content and inclusion
bolt heads made of material 1 presents clear evidence that the cur- types that are deleterious to processing performance and nal prod-
rent rules of thumb are neither effective nor adequate in fullling uct properties.
the needs of the new design developments in the rapidly expand- Dodd and Bai (1987) indicate that the inclusions refer to non-
ing CH industry. They do not provide enough guidance to fulll the metallic particles that are held mechanically in the material matrix.
needs of new design developments in the CH industry. Therefore, Such particles include oxides, suldes or silicates. The presence of
the CH industry is in need of systematic and theoretical studies to these inclusions strongly inuences the ductile failure mechanism.
understand the failure mechanisms and the different parameters Thus control of cleanliness and second-phase particle shape can
contributing to failure initiation and propagation in cold-headed inuence material ductility.
products. The inclusion rating process for this study was performed using
These examples of failed cold-headed products reveal the appar- automatic image analysis inclusion ratings in accordance with
ent lack of knowledge about the internal ductile failure mechanism microscopic method A (worst elds) of E45 of the ASTM standard.
and the role of the ASB phenomenon in the presence of these defects Worst eld is a class of rating in which the specimen is rated for
in cold-headed products each type of inclusion by assigning the value of the highest severity
rating observed of the inclusion type anywhere on the specimen
surface.
2. Experimental
Method A requires a survey of a polished surface square area
of 160 mm2 at 100 magnication parallel to the longitudinal axis
2.1. Material selection and characterization
of the wire. The severity level of the worst elds and the series
type of four inclusion types (A, B, C, and D) should be reported for
The CH quality AISI steel grades chosen for this research are:
every specimen depending on Tables 2 and 3. Table 2 presents the
1008, 1018, 1038, 1541, 8640 and dual phase (DP) steel which cover
low- and medium-carbon content steels and also provide different
CH quality steels ranging from best to poorest CH formability steels.
The 1018, 1038, 1541, and 8640 steels were received in rod form
from Ivaco Rolling Mills (Ontario, Canada), while the 1008 steel was
provided from the undeformed part of the bolts made of material
1.
In addition, DP steel was also included in the test materials as it
possesses unique properties which show great potential in the CH
industry. Pierson (1986) reported that the DP steel is intended to
be cold headed directly from the green rod and reaches the nal
properties of the cold-headed products without heat treatment.
Optical microscopy revealed a microstructure consisting of fer-
rite matrix and lamellar pearlite for all steel grades except the DP
steel which consisted predominately of ferrite grains and a marten-
sitic second phase. The as-received microstructure of all the testing
materials is shown in Fig. 4. Table 1 lists the chemical composition
of the selected material.
Quasi-static tensile tests were performed on specimens Fig. 5. The tensile quasi-static true-stress vs. true strain curves for the testing mate-
machined to ASTM-E8 standard at an approximate strain rate rials.
 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311 4297

Table 2
Total inclusion length or number for minimum severity level numbers (method A) (ASTM E45-97).

Severity level Total length in one eld at100, min. (mm) Number of inclusions in one eld

Type A Type B Type C Type D

0.5 3.7 1.7 1.8 1

1 12.7 7.7 7.6 4
1.5 26.1 18.4 17.6 9
2 43.6 34.3 32.0 16

Table 3 Table 4
Inclusion width and diameter parameters (method A) (ASTM E45-97). Inclusion rating test results for each testing material using ASTM E45-97 method
A.a .
Inclusion type Thin series (T) Heavy series (H)
AISI grade A B C D
Width min. Width max. Width min. Width max.
(m) (m) (m) (m) 1008 None None None D1
steel None None None D1
Type A 2 4 >4 12
Type B 2 9 >9 15 None None C1 D1
Type C 2 5 >5 12 1018
None None None D1
Type D 2 8 >8 13 steel
None None None D1

1038 None None C1 D1

minimum severity level number for each inclusion type and Table 3 steel A1 None None D1
provides the inclusion dimensions for the thin (T) and heavy (H) A1 None C2 D1
series according to method A. 1541
A1 None None D1
steel
This microscopic method places inclusions into several A1 None None D1
composition-related categories: suldes, oxides, and silicates. The A1 None None D1
8640
typical chemical types of inclusions in Type A are sulde stringers, A1 None C2 D1H
steel
while Type C are silicate stringers. Type B are oxide stringers while None None None D1H
Type D is a globular oxide. Method A requires length measurement None None C1 D1
of Type A inclusions, the total stringer length of Type B and C inclu- DP None None C2 D1
sions, and the number of Type D inclusions. Type D globular oxide steel None None C1 D1
A1 None None D1
may not exceed an aspect ratio of 5:1.
a
The inclusion rating test results for the six testing materials are Cleanliness tests were performed by Ivaco Rolling Mills, LOrignal, Ontario.
displayed in Table 4 and the morphologies of the inclusion types
are shown in Figs. 68.

Fig. 6. Inclusion type A1 in the testing materials.

4298 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311
Fig. 7. Inclusion type C1 in the testing materials.
Zhang and Thomas (2002) reported that the inclusion size is
very important because large inclusions are the most harmful to the
mechanical properties of CH materials. Sometimes a catastrophic
defect is caused by a single large inclusion in an entire steel heat.
Thus, producing clean steel requires not only controlling the mean
inclusion content in the steel but also avoiding inclusions larger
than the critical size.
2.3. Tests
The drop weight compression test (DWCT), as described by
Nickoletopoulos (2000), facilitates the testing of the CHP and
is capable of generating ASBs during upset testing. The DWCT
machine, as illustrated in Fig. 9, consists of a tower enabling inter-
changeable weight plates to be dropped from heights up to 2.4 m.
The die set conguration (Fig. 10) rests on a central column that is
welded to the base of the DWCT machine. Specimens for CH were
machined with a tolerance of 0.02 mm from as-rolled rod material
to a cylindrical conguration of 5.3 mm in diameter with aspect
ratios of 1.6 and 1.8. A series of tests is performed by varying the
weights until internal fracture is found at different locations inside
the ASB.
A guided cylindrical-pocket die set was designed for the DWCT
machine. A sleeve guides the dies and reduces die movement during
testing. Air vents in the sleeve prevent pressure build-up. An impact
plate is used on top of the upper die to prevent direct impact with
the die. Fig. 10 presents a schematic assembly of a test specimen
which is placed between the upper and lower dies. This assembly
rests atop the force sensor inside a cylindrical pocket.
2.4. Sample preparation
The cold-headed samples were sectioned transverse to the head-
ing direction using a high-concentration diamond wafering blade
on a slow speed (200 rpm) cut off saw that was operated with an
oil coolant. Sectioned specimens were mounted in Bakelite and
prepared for metallographic examination using automated tech-
niques for grinding and polishing. Specically, the specimens were
ground with successive papers of SiC from 280 grit to 600, fol-
lowed by rough polishing using a 9 m diamond suspension with
 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311
Fig. 8. Inclusion type D1 in the testing materials.
an alcohol based lubricant on a silk cloth and nal polishing with 3
and 1 m on a nylon and porous pad, respectively. After ultrasonic
cleaning, the specimens were etched by immersion in a 2% nital
solution.
Microstructural examination was performed using optical light
microscopy with polarizing capabilities and image analysis soft-
ware to examine the overall characteristics of the ASBs and the
voids and cracks initiated in the interior or along the ASBs. High-
resolution imaging for examining the topography of the voids
and cracks in the ASBs was performed using scanning elec-
tron microscopy operated in secondary electron (SE) mode and
equipped with an X-ray detector (EDAX) for elemental analysis of
secondary phases. Mapping the topography of the voids and cracks
in the ASBs was performed using an Olympus OLS1200 laser scan-
ning confocal microscope (LSCM) equipped with an argon laser
(488 nm) and objective specications of up to 100 with a 6 opti-
cal zoom. As the image produced by scanning the x- and y-axes
with a spot laser beam in the LSCM originates from a shallow focus
depth, for the DP steel specimens a series of precisely focused opti-
4299
cal images were acquired along the z-axis of the microscope. The
successive images were then overlapped using specialized recon-
struction software to obtain a three-dimensional extended focus
image of the cross-sectional view that contained both height and
intensity information. Specically, the topographic features in the
extended focus image of the ASB structure enabled the characteri-
zation of the size of the width and depth of the voids and cracks.
3. Results and discussion
In general, microscopic examination and FE analysis of the DWCT
specimens at various regions of each specimen during different
deformation levels showed that the deformation concentrates in
the ASBs (Sabih et al., 2005, 2006a,b). The microstructure of the
DASBs is laminar consisting of highly elongated grain layers of fer-
rite and pearlite or second-phase particles (Fig. 11). Voids and cracks
found along the ASBs inside DWCT specimens (Fig. 12) coincided
with the cracking that occurs inside the head of the industrial bolt
shown in Fig. 1. This nding agrees with that of Qiang and Bassim
4300 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311
Fig. 9. Schematic of the DWCT machine.
Fig. 10. DWCT specimen inside the die set arrangement.
Fig. 11. The microstructure of the DASB (1541 steel, aspect ratio of 1.6, drop height
of 2.4 m and drop weight 29.5 kg).
(2003) who studied the effect of process parameters on the forma-
tion of the ASBs in metals and observed the initiation and growth of
cracks along the ASBs. Similar observation was reported in the work
of Nemat-Nasser et al. (2001) who reported in their studies of ASB
Fig. 12. Microscopic views showing the internal cracks and/or voids along the ASB
(DP steel, aspect ratio = 1.8, dropped weight = 33.8 kg, height = 2.4 m).
phenomenon and the associated failure that these bands usually
precede the failure mechanism for most metals and alloys.
Ductile fracture occurs progressively by void nucleation and
growth mechanisms with increasing plastic strain. Different mech-
anisms of ductile failure have been reviewed in many works: Dodd
and Bai (1987), Dodd and Atkins (1983) and Garrison and Moody
(1987). All these reviews consider the most important factor affect-
ing the material formability is the number and morphology of
inclusions and second-phase particles.
Dodd and Atkins (1983) found in their study of ow localization
in shear deformation of void containing and void-free solids that
ductile failure occurs following this sequence of events:
(1) Nucleation of voids occurs by decohesion at second-phase parti-
cles or at inclusionmatrix interfaces, or by particle or inclusion
cracking. Dodd and Hartley (1991) concluded that in certain
cases, these voids can nucleate at grain boundaries or grain
boundary triple points.
(2) An increase in the volume fraction of voids with increasing
applied plastic strain.
(3) At a critical strain, plastic deformation is localized along shear
bands between the voids.
(4) Cracks occur along the bands of intense shear, forming what
Rogers (1959) calls void sheets mechanism of crack propaga-
tion.
Masheshwari et al. (1978) studied quality requirements for steel
CH grades and concluded that the second-phase particles and inclu-
sions, such as heavy silicate, sulphide, alumina and globular oxide
inclusions, should not exceed the acceptable limits of the cold head-
ing material. Thomson (1990) pointed out that these particles and
inclusions act as stress raisers and can support microvoid nucle-
ation, growth, and coalescence, leading to ductile fracture.
Meyers et al. (2001) reported that the localization of the ASBs
is an important deformation mode that causes ductile fracture in
metals. Wei et al. (2006) concluded in their study of mechanical
behavior and dynamic failure under uniaxial compression that the
thermal and geometric softening and material ow patterns are
responsible for dynamic failure under compressive loads.
Ductile failure caused by ASBs within the DWCT specimens
under compressive loads is a complicated process which depends
on the interplay of many factors. The following sections will detail
the ductile failure mechanisms, the void nucleation, coalescence
and crack initiation mechanism. Furthermore, the interplay of
inclusion types and content, material microstructure, material ow
pattern and the localization phenomenon will be discussed along
with the impact of these factors on the occurrence of ductile failure.
3.1. Void initiation mechanism
A careful reading of the literature shows that there is a debate
about the issue of void nucleation under a compressive state of
stress. Petersen et al. (1997) reported that ductile failure is unlikely
to occur under a compressive stress state. Teng et al. (2007) stud-
ied the transition from adiabatic shear banding to fracture and
reported that ductile fracture under high-negative stress triaxial-
ity cannot occur. On the other hand, many researchers have found
that inclusions or inter-metallic particles serve as void nucleation
sites under compressive loads: Puttick (1959) showed in his study
of ductile fracture that microvoid initiation and growth is caused
by the debonding of second-phase particles produced by deforma-
tion in the surrounding matrix. Cox and Low (1974) found that void
nucleation and growth occurs more readily at the larger inclusions.
Rogers (1960) explained the role of the small and large particles
and inclusion in the initiation and growth of voids leading to a duc-
tile fracture. Similarly, Gurland (1972) also explained the effect of
 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311 4301

Fig. 13. DWCT specimen of 1018 steel showing the locations of the CASB, UDZ, and LDZ. Solid lines represent imaginary boundaries of the ASB and the UDZ and LDZ. Big arrows
represent the shear direction and the small arrow shows the material ow pattern. The dashed arrow represents the mid of the CASB (aspect ratio = 1.8, drop height = 2.4 m,
drop weight 22.5 kg).

the inclusions on the nucleation, growth, and coalescence of micro-
scopic voids causing fracture of ductile solids. Horstemeyer and
Gokhale (1999) modeled voidcrack nucleation process for duc-
tile metals with second-phase particles. They concluded from the
modeled compression, tension, and torsion experiments that the
voidcrack nucleation can take place under compressive loading
condition.
Nemat-Nasser et al. (2001) also reported that cracks of this kind,
which are produced by local defects, can be created by a local tensile
stress state under an overall compressive applied stress. They also
found that the occurrence of tensile stresses in the vicinity of the
inclusions and/or second-phase particles create the classical con-
ditions for void nucleation and growth. Thus, fracture initiation in
ASBs under compressive loading of the DWCTs is highly probable.
The present ndings are supported by previous work by Dodd
and Atkins (1983) showing that void growth is sensitive to a
hydrostatic stress state. Furthermore, they found that voids nucle-
ate at second-phase particles or inclusions within the ASB under
compressive hydrostatic stresses. Under compressive loading con-
ditions with a local tensile stress state, both void nucleation and
growth can occur in steels, albeit at a lower rate in comparison to
the dominating failure mechanisms of void formation and propaga-
tion under tensile loading. Park and Thompson (1988) have shown
that void nucleation occurs at both sides of the tensile pole of a
particle under compressive loads. Horstemeyer and Gokhale (1999)
reported the possible occurrence of void nucleation even under
compression loads as tensile states exist for compression loadings.
3.2. Effect of material ow pattern on void nucleation
Void nucleation around second-phase particles and non-
metallic inclusions is highly inuenced by the nature of the material
ow during deformation. Sarruf et al. (1999) reported that the metal
ow pattern during the CH process supports the void nucleation
mechanism at non-metallic inclusions and the surrounding mate-
rial. Meyers et al. (2001) reported that the material ow pattern
during shear localization is an important factor in sub-grain break-
age and rotation. Tvergaard (1981) reported that this mechanism is
responsible for void nucleation due to brittle cracking or decoher-
ence of inclusions.
Detailed microscopic examination of various regions of the
DWCT specimens showed that the deformation is concentrated in
the localized central ASBs (CASB) with well-dened ow contours
starting from the upper dead zone (UDZ) and the lower dead zone
(LDZ), as illustrated in Fig. 13. Specically, there is a sharp transi-
tion in the behavior of the material ow as observed by the contour
lines which traverse from the dead zones to the CASB (Fig. 14(a)), to
the CASB edge at the mid-axis (Fig. 14(b) and (c)). In the deformed
region within the DWCT dies, the material ow behavior showed
elongated semi-elliptical contours that increasingly concentrated

Fig. 14. Material ow contours for various regions on a DWCT specimen (drop weight = 35.5 kg and height = 2.4 m) showing: (a) the dead zone region adjacent to the ASB
located at the center of the DWCT specimen (b) the region adjacent to the UDZ, (c) the region near the specimen edge at the mid-axis, (d) the region deformed within the
DWCT dies (25), (e) the region deformed within the DWCT dies (100), (f) the region in the CASB showing the merging of two dead zones and (g) the center of the CASB.
4302 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311
Fig. 15. Excessive deformation at the region between the UDZ, the LDZ and the
CASB resulting in microscopic crack formation at the boundary between the top and
bottom dead metal zones and at the inection point of material ow (DP steel, aspect
ratio of 1.6, drop height of 2.4 m and drop weight 31.6 kg).
towards the mid-axis of the specimen, as shown in Fig. 14(d) and (e).
At the center of the DWCT specimen, the CASB zone was observed
to have highly deformed material with voids (Fig. 14(f) and (g)).
With further deformation, the region extended along (f), (g), (b)
and (c), and was observed to have excessive deformation result-
ing in microscopic crack formation at the boundary between the
(UDZ) and the (LDZ) and at the inection point of material ow
(Fig. 15).
An analysis of the void characteristics, using the LSCM, indi-
cated that an increase in the plastic strain caused both an increase
in the void density and size, as illustrated in Fig. 16. Thus, in
the CASB region, longer cracks were observed with increasing
drop weight for the DWCT specimens. The occurrence of the long
cracks in the CASB was usually observed in two zones. The rst
zone was along the center of the CASB (Fig. 14(f)), and the other
Fig. 16. A comparison of void density inside the ASB of two specimens (DP steel,
aspect ratio of 1.8 and drop height of 2.4 m) that were deformed with a drop weight of
(a) 38 kg and (b) 31.5 kg. The dark heights on the surface represent the void locations
and their inverted depth.
was found in the upper and lower regions adjacent to the CASB
(Fig. 14(a)(c)).
3.3. Void nucleation mechanism at the inclusionsmatrix
interface
Siruguet and Leblond (2004a,b) concluded that the void nucle-
ation, growth and coalescence mechanism is highly dependant on
the stress state under which deformation takes place. Fig. 17 dis-
plays possible void nucleation mechanisms under different stress
states in order to have a better understanding of the ductile fail-
ure occurring inside the ASB. When the stress triaxiality is high
(tensile loads/positive), voids grow in the directions of the tensile
loads (Fig. 17(b) and (c)), and the enclosed inclusions do not inu-
ence their growth. Moreover, it has been reported by Pardoen and
Delannay (1998) that the initially spherical voids will change to
ellipsoidal voids under tensile triaxiality.
Siruguet and Leblond (2004a,b) showed that under compres-
sive loads from all directions, or negative stress triaxiality, voids
can be subjected to compression in one or several directions,
thereby changing their shape. Also, if they are still in contact
with the enclosed inclusions in these directions, inclusions pre-
vent their shrinkage, a process known as void locking illustrated in
Fig. 17(d). Pardoen and Delannay (1998) reported that under low-
compressive stress triaxiality, the ductile matrix may ow around
the inclusion resulting in the failure of the interface forming small
voids at the sides of the inclusion (Fig. 17(e)). Conversely, under
high-compressive loads, the fracture of the inclusion is possible
(Fig. 17(f)).
In the case of shear banding under compressive loads, the forma-
tion of microvoids has been reported by Pirondi and Bonora (2003)
and Sabih et al. (2005, 2006a). They indicated that the microvoids
have smaller dimensions than the tensile cases. They also con-
cluded that voids under compressive loads do not increase their
dimensions but stretch and eventually coalesce by internal necking,
known as microvoid sheeting (Fig. 17(g)).
Evidence of void and crack initiation and propagation at the non-
metallic inclusions and stringers sites was put forward in the work
of Vora and Polonis (1976). They found that the weakly bonded
particles to the matrix had elongated in the direction of material
ow. Once decohesion between the matrix and the inclusion or
second-phase particle occurred, the nucleated voids were observed
to grow with increasing strain from an initial spherical shape to an
ellipsoidal form or sheet form, as illustrated in Fig. 17(h). Boyer et
al. (2002) modeled void growth under different stress states and
concluded that void volume increases under compressive loads if
these voids are lled with inclusions.
Meyers et al. (2001) noticed that the different rates of material
ow around the inclusion supports decohesion and void nucleation
mechanisms between the matrix and the inclusion or the second-
phase particle as shown in Fig. 17(i).
In DWC testing, the deformation takes place under compressive
stress triaxiality over almost the entire specimen. The ASB occurs
inside the DWCT as a result of the softening mechanisms and the
shear deformation between the edges and the center of the band.
The combined compressive and shearing deformation results in
ductile failure inside the band. The void nucleation and coalescence
follows some of the illustrated mechanisms in Figs. 17 and 18.
Fig. 18 displays some metallographic observations of the void
nucleation mechanisms around the spherical inclusions (Type D).
Fig. 20(b)(f) show the steps of the void sheeting process inside the
CASB of the 1541 steel. This process starts when a void nucleates at
the edges of the inclusion due to the material ow around it. With
more deformation, the inclusion changes its shape from spheri-
cal to ellipsoidal and the void propagates along the material ow
direction forming a void sheet (Fig. 18(e)). This void sheet occurs
 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311 4303

Fig. 17. Illustrations of the void and crack nucleation mechanisms around the inclusions under different stress states: (a) inclusion inside the undeformed material (no void);
(b) under multiaxial tensile loads, void growth in all directions; (c) under uniaxial tensile load, void nucleation in the load direction; (d) under multiaxial compressive loads,
void shrinks on the inclusion (void locking); (e) under uniaxial compressive load, void nucleation at the sides of the inclusion; (f) breakage of inclusion under multiaxial
compressive loads; (g) void sheet nucleation and growth at the sides of the inclusion under compressive and shear loads; (h) fractured or elongated weak inclusion along
the material ow direction under combined compression and shear stresses; (i) decohesion between the inclusion and the matrix due to rotation of inclusions caused by
different rates of material ow around the inclusion.

in a manner similar to the void sheeting mechanism illustrated in
Fig. 17(h).
The metallographic inspection showed that some of the small
voids change their shape to an ellipsoidal geometry, while the large
inclusions were not affected by material deformation and contin-
ued to remain in their spherical shape. However, it was noticed that
void sheets initiated at the vicinity of the big inclusions and propa-
gated along the material ow direction (Fig. 18(g)). These ndings
are supported by the results of the ductile damage nite element
model used by Boyer et al. (2002) to predict void growth under
normal mean stress.
Due to the different rates of material ow around the inclusions,
the big spherical inclusion rotates in the direction of the high-ow
rate. This results in decohesion at the interface between the inclu-
sion and the surrounding matrix. These ndings are supported by
similar observations made by Meyers et al. (2001), who reported
that the material ow pattern during shear localization is an impor-
tant factor in the breakage and rotation of inclusions.
Sabih et al. (2005, 2006a) showed in their metallographic inves-
tigations and FE simulations that at low-plastic strains (less than 1)
and low-negative stress triaxiality, voids were observed to initiate
around the elongated inclusions (Fig. 19(b)). Bandstra et al. (2004)
reported void nucleation at the non-metallic inclusionmatrix
interface at small strains. In this work, inclusions were also
observed to fracture during deformation (Fig. 19(c)). According to
the microscopical examination of fracture surfaces made by Broek
(1973), this can cause stress concentrations promoting ductile frac-
ture.
Type A and C (long inclusion) stringers are the favored places for
void nucleation and growth within the ASBs. The metallographic
study of the DWCT specimens showed that the void sheeting pro-
cess took place by decohesion at the stringermatrix interface in
several steps. The void sheeting mechanism around the stringers
is the result of the material ow and strain localization within the
ASB. As shown in Fig. 14, the stringers move along the material ow
direction and change their alignment by approximately 90 . This
sharp change in the ow direction at the early stages of deformation
triggers the decohesion process at the stringermatrix interface at
the boundaries of the CASB (Fig. 14(b) and (c)).
When these stringers join the localized CASB, the softening
mechanisms and the high-strain rate deformation within the band
support the void sheeting process along the stringers sides and
around the spherical inclusions (Fig. 14(f) and (g)). Fig. 20(a) dis-
plays the sharp change in the stringers direction which supports
4304 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311

Fig. 18. Changes in the inclusion morphology and void sheet nucleation mechanism around inclusions at different deformation levels in 1541 steel: (a) spherical inclusion
inside the undeformed material, (b) inclusion shape changed from a spherical to an ellipsoidal form inside the DASB, (c) elongated inclusion inside the DASB, (d) void sheet
initiated at the edges of the elongated inclusion, (e) void sheet growth inside the DASB, (f) a large number of void sheets inside the highly deformed TASB, (g) void sheet
propagating ahead of the inclusion, and (h) rotation of the inclusion.
 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311 4305

Fig. 19. The morphology of inclusions: (a) spherical inclusion in as-received (undeformed) DP steel, (b) elongated particles with adjacent voids initiated at low-plastic strains,
and (c) fractured elliptical inclusion.

Fig. 20. Void sheet mechanism around the stringers: (a and b) 1038 steel, aspect ratio of 1.8, drop height of 2.4 m and drop weight 26.2 kg, (c) 1018 steel, aspect ratio of 1.6,
drop height of 2.4 m and drop weight 28.5 kg, (d) 8640 steel, aspect ratio of 1.6, drop height of 2.4 m and drop weight 29.5 kg, and (e) 1541 steel, aspect ratio of 1.6, drop height
of 2.4 m and drop weight 23.4 kg.
4306 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311

the void sheets inside the DWCT of 1541 steel. Usually, the voids
nucleate around the stringer, forming the void sheet inside the
deformed region. With further deformation, the void sheet, which
is considered as another softening mechanism, propagates along
the localized CASB (Fig. 20(a) and (b)).
In addition to void sheeting around the stringers within the
CASB, metallographic inspection revealed that these stringers are
the source of ductile failure inside the region deformed within the
DWCT dies (Fig. 20(a) and (c)). In this region, the material ow pat-
tern forces the stringers to take the shape of semi-elliptical contours
that become increasingly concentrated towards the mid-axis of the
specimen: Fig. 20(a), (d) and (e). The stringers and the surrounding
material at these locations deform under low-compressive stress
triaxiality in comparison with the center of the CASB. Under this
stress state, and with further deformation, cracks initiate at the tip
of the semi-elliptical stringers due to the stretch-like deformation.
This stretching occurs because the ow of the semi-elliptical sides
is restricted, due to the friction with the die faces, while the mate-
rial continues to push the tip of the semi-elliptical stringers away
from the specimen center.
3.4. Void nucleation mechanism at the grain boundaries and the
second phasematrix interface
As presented in Section 2.1, the materials tested are
ferriticpearlitic steels, except for the DP steel which has a
ferritemartensite microstructure. The pearlite grains consist of
cementite lamella in a ferrite matrix. This type of microstructure
makes the materials subject to a new source of void nucleation
during plastic deformation under compressive loading.

Fig. 21. SEM micrographs display the void nucleation at the ferritecementite interface, at the grain boundaries and around the inclusions inside the DASB: (a) lamellar
pearlite (ferritecementite) in the undeformed zone (b) elongated pearlite in the direction of material ow, (c) void nucleation at the ferritecementite interface, (d) void
nucleation at the grain boundaries, (e) void nucleation at the ferritecementite interface and at the grain boundaries, and (f) void nucleation around an elongated inclusion
(1038 steel, aspect ratio of 1.8, drop height of 2.4 m and drop weight 33 kg).
 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311 4307

Table 5
Inclusion rating result with the minimum inclusion length or number according to the minimum severity level numbers (method A) (ASTM E45-97).

AISI steel grade A1 Min. length (mm) C1 Min. length (mm) C2 Min. length (mm) D1 Min. number

1008 None None None D1 4

1018 None C1 7.6 None D1 4
1038 A1 12.7 C1 7.6 None D1 4
1541 A1 12.7 C1 7.6 C2 32 D1 4
8640 A1 12.7 C1 7.6 C2 32 D1, D1H 4
DP A1 12.7 C1 7.6 C2 32 D1 4

D1H is a heavy series of Type D1 inclusion, the inclusion width range is between 8 and 14 m. The minimum width of the other inclusion types is 2 m.

Cox and Low (1974) referred in their study of ductile fracture of
AISI 4340 steel that this failure is caused by void sheets composed
of small voids nucleated at the matrixcementite (second-phase
particles) interface. Gurland (1972) studied the formation of cracks
and voids at second-phase particles (carbides) under tension, com-
pression and torsion. Hahn and Roseneld (1975) reported that void
growth and coalescence was observed the cracked inclusions and
the grain boundaries. Nieh and Nix (1980) explained void growth
and coalescence at grain boundaries. In conclusion, non-metallic
inclusions, second-phase particles, and grain boundaries are con-
sidered to be the best candidate sites for void nucleation and ductile
failure.
Tanguy et al. (2006) have also reported that the presence of
second-phase particles, such as lamellar cementite or martensite,
are a source of void nucleation. Chen et al. (2002) showed that
lamellar cementite has a delayed deformation rate as compared to
ferrite. Valiente et al. (2005) explained that the lamellar cementite
of a pearlitic grain has a much higher stiffness than ferrite.
Valiente et al. (2005) showed that due to microstructural
aspects, the difference in the local ow behavior can lead to cemen-
tite elongation and fragmentation. This triggers void nucleation at
the discontinuities produced by the breakage of cementite and by
decohesion of atomic bonds at the interface between the matrix
and inclusion or second-phase particles. The voids grow, assisted
by plastic deformation, until they link up. Fragmentation of cemen-
tite within ASBs has also been reported by Chen et al. (2003). Argon
(1976) found that voids were more easily nucleated at the larger
cementite particles than at the smaller ones.
Beyond the early stages of deformation in DWCT specimens,
metallographic study by SEM microscopy showed that the voids
nucleated at the cementiteferrite interface, the pearlitematrix
interface and at grain boundaries (Fig. 21).
Within the matrix, the cementite inside the pearlite grains far
off the shear band have a lamellar shape as shown in Fig. 21(a).
Lamellar cementite inside the pearlite on the shear band side is
elongated along the material ow direction (Fig. 21(b)). At the
high-deformation zone close to the CASB, broken lamellar cemen-
tite was noticed, and clear void nucleation took place at the
ferritecementite interface (Fig. 21(c)).
Voids and microcracks nucleated at ferritepearlite interfaces
close to the localized plastic deformation zone of the CASB
(Fig. 21(d)). Closer to the highly localized plastic deformation within
the CASB, more void nucleation and cracking along the grain bound-
aries along the ASB was observed (Fig. 21(e)). These voids and
microcracks, along with the voids around the inclusions (Fig. 21(f))
contribute to ductile cracking along the ASBs.
3.5. Ductile failure characterization in the testing materials
The metallographic study of the DWCT specimens showed that
all materials experienced ductile failure by void nucleation and
coalescence, forming cracks along the CASBs. The ductile failure
of each material was the result of the contribution of all the
mechanisms of void nucleation at the inclusionmatrix interface,
second phasematrix interface and at the grain boundaries, as
previously discussed. However, the level of contribution of each
mechanism in the nal ductile failure varied depending on mate-
rial.
Ductile failure along the ASB inside the DWCT occurred under
compressive loads. The voids and cracks found in the failed zones
have the narrow, stretched, elongated shape or the void sheet shape
along the highly localized regions of the ASB.
In general, it is possible to classify the ductile failure into two
groups. The rst group consists of materials that experienced very

Fig. 22. Ductile failure in DWCT specimens of 1008 steel: (a) a long crack along the CASB (aspect ratio of 1.6, and drop weight 37.6 kg), (b) void nucleation in the DASB (aspect
ratio of 1.6, and drop weight 26.2 kg).
4308 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311

large cracks, including 1008 steel and 1018 steel. The second group
consists of the material that experienced short ductile cracking
along the ASB. This group includes 1038 steel, 1541 steel, 8640 steel
and DP steel.
The rst group is composed of soft steels with low-inclusion
content. Table 5 shows that the only inclusion type in the unde-
formed 1008 steel is the spherical inclusions (Type D1), while the
undeformed 1018 steel contains the long stringer or inclusion Type
C1 in addition to Type D1 inclusions.
Both steels experienced very long cracks which extended for
more than 3 mm along the mid-axis of the CASB, as shown in
Figs. 22(a) and 23(a) for 1008 steel and 1018 steel, respectively. The
SEM metallographic inspection of the cracked specimens revealed
that void nucleation and coalescence took place at the boundaries
of the highly elongated grain within the CASB in contribution to the
microcracks at the ferritecementite interface. The metallographic
investigation of the role of the inclusions type in the 1008 steel
(inclusion Type D1) showed that voids nucleated along the inclu-
sions leading to shorter cracks in comparison to the crack initiated
at the grain boundaries (Fig. 22(b)).
Similar conclusions can be made with respect to the 1018 steel
DWCT specimens. The metallographic investigation showed that all
the void nucleation mechanisms contributed to the ductile cracks
found within the highly localized deformed CASB in the DWCT
specimens made of 1018 steel (Fig. 23). Void nucleation took place
around the inclusions (Fig. 23(b)), and at the interface between
both types of inclusions (Type C1 and Type D1) and the matrix.
The microscopic observation revealed that at the early stages of
deformation, void sheets and short cracks extend along the grain
boundaries in addition to microcracks at the ferritecementite
interface. With further deformation, the coalescence of these void
sheets and cracks form bigger cracks that extend over 3 mm along
the center of the CASB (Fig. 23(a)). The void nucleation along the
Type C1 inclusion is readily apparent in Fig. 20(c).
The materials in the second group can be characterized by the
occurrence of the TASBs that are accompanied with short cracks.
These cracks were noticed to be within, or at, the sides of the TASB
zone and the DASB zone of the band. As in the rst group, all of
the void nucleation mechanisms contributed to the ductile cracks
found within the highly localized deformed and the transformed
part of the CASB. However, the long inclusions (Type A1, C1 and C2)
played an important role in ductile failure initiation as previously
discussed. The presence of inclusion type C2 in 1541 steel, 8640
steel and DP steel means that stringers with a total minimum length
of 32 mm are present in the matrix. As discussed previously, this
long inclusion is the favored location of void nucleation by the deco-
hesion mechanism at the inclusionmatrix interface. With further
straining inside the localized ASB, these voids coalesce around the
stretched long inclusion inside the band. Figs. 20(a), 20(b) and 24(a)
display the long cracks which initiated within the CASB of the 1038

Fig. 23. Ductile failure in the DWCT specimens of 1018 steel: (a) a long crack along the CASB (aspect ratio of 1.6, and drop weight 37.6 kg), (b) cracks at the grain boundaries
(aspect ratio of 1.8, and drop weight 22.5 kg), and (c) void nucleation around the inclusion (aspect ratio of 1.6, and drop weight 20.2 kg).
 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311 4309

Fig. 24. Crack initiation along the TASB: (a) 1038 steel, aspect ratio of 1.8, drop height of 2.4 m and drop weight 33 kg, (b) 8640 steel, aspect ratio of 1.8, drop height of 2.4 m
and drop weight 29.5 kg.

Fig. 25. Void coalescence in the ASB region of deformed DWCT specimen of DP steel: (a) 500 and (b) 1000 (aspect ratio of 1.8, drop height of 2.4 m and drop weight
35.6 kg).

steel, and shows the severe change of the material ow pattern.
This is known to cause the breakage of the long inclusion, provid-
ing another means of void sheet nucleation. Fig. 24(b) shows that
voids nucleated at the interface of the long inclusionmatrix in the
8640 steel. Fig. 25 shows ductile failure along the ASB inside the
DWCT specimen of DP steel. The ductile cracking mechanisms in
the 1541 steel are shown in Fig. 18.
Careful investigation of the microstructure at the side of the
CASB of the second group materials reveals that the elongated fer-
rite and pearlite grains also helped in the nucleation of narrow and
stretched void sheets along the localized band. Nevertheless, the
role of the long inclusion clearly plays a much larger part in ductile
failure.
4. Conclusion
The occurrence of internal ductile failure in cold-headed prod-
ucts presents a major obstacle in the fast expanding CH industry.
The current design rules of thumb failed in avoiding this type of
failure which results in reducing the materials capability to carry
loads. This internal failure may lead to catastrophic brittle fracture
under tensile loads despite the ductile nature of the material.
The lack of specialized studies regarding the mechanisms of
internal ductile failure in cold-headed products and the failure of
the current CH design procedures to fulll the needs of new devel-
opments in the expanding CH industry motivated this research.
A comprehensive testing and investigation methodology was
4310 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311
used to reveal the complicated interplay of process and material
parameters contributing to the initiation and propagation of the
internal ductile failure in six CH quality AISI steel grades.
The metallurgical and microscopic studies showed that inter-
nal ductile failure occurs progressively by void nucleation and
growth mechanisms with increasing plastic strain inside the highly
localized CASBs. The most important factor affecting material
formability is the number and morphology of any inclusions and
second-phase particles. Moreover, the thermal and geometric soft-
ening mechanisms and material ow pattern are responsible for
dynamic failure under compressive loads.
It was conrmed that ductile failure occurs by a void nucleation
mechanism which occurs by decohesion at second-phase particles
and/or at inclusionmatrix interfaces, by particle or inclusion crack-
ing, or at grain boundaries. The localization process inside the CASBs
supported the void growth and coalescence along the band form-
ing narrow void sheets. Under compressive loading conditions, the
void nucleation, growth and coalescence mechanism occurred due
to the local tensile stress state in the vicinity of the inclusions and/or
second-phase particles.
The metallurgical investigations showed that the nature of the
metal ow pattern during the CH process, and the different rates of
material ow around the inclusions, support decohesion and void
nucleation mechanisms between the matrix and the inclusion or
the second-phase particle in the highly localize CASBs.
The void sheeting process starts when a void nucleates at the
edge of the inclusion due to the material ow around it. With more
deformation, the inclusion changes its shape from spherical to ellip-
soidal and the void propagates along the material ow direction
forming a void sheet.
Under combined compressive and shearing deformation, the
large inclusions continued to retain their spherical shape. However,
it was noticed that void sheets initiated in the vicinity of the large
inclusions and propagated along the material ow direction. Type A
and C long inclusions are the favored places for void nucleation and
growth within the CASBs. The void sheeting process took place by
decohesion at the stringermatrix interface in several steps. The
sharp change in the ow direction at the early stages of defor-
mation triggered the decohesion process at the stringermatrix
interface. The thermal softening and the high-strain rate deforma-
tion within the localizing band support the void sheeting process
along the stringers sides and around the spherical inclusions. With
further deformation, the void sheet propagates along the localized
CASB.
In addition to non-metallic inclusions, the second-phase parti-
cles and grain boundaries are considered to be the best candidate
sites for void nucleation and ductile failure. The difference in the
local ow behavior, due to microstructural aspects, can lead to
cementite elongation and fragmentation. This triggers void nucle-
ation at the discontinuities produced by the breakage of cementite
and by decohesion at the interface between the matrix and inclu-
sion or second-phase particles.
The microstructure of the DASBs is laminar, consisting of narrow,
stretched, and highly elongated grain layers of ferrite and pearlite
or second-phase particles. Void sheets and cracks are found in the
failed zones along the ASBs.
All materials tested experienced ductile failure by void nucle-
ation and coalescence, forming cracks along the ASBs. The ductile
failure of each testing material was the result of the contribution
of all the mechanisms of void nucleation at the inclusionmatrix
interface, second phasematrix interface and at the grain bound-
aries. However, the level of contribution of each mechanism in the
nal ductile failure varied depending on the material.
In summary, a comprehensive experimental and metallurgical
study of the ASB phenomenon and the associated ductile failure
were introduced in this study to uncover the complex interplay of
different material and process parameters controlling the failure
mechanisms within the ASBs in cold heading process.
The ductile failure initiation under compressive loads is a
debated subject of many works in the literature. However, this
detailed study reached important ndings that uncovered the
complex interplay of the different mechanisms of ductile failure
initiation and growth under compressive loads. All these ndings
were based on detailed supporting experimental and metallurgical
evidences about these mechanisms.
References
Argon, A.S., 1976. Formation of cavities from non-deformable second-phase particles
in low temperature ductile fracture. J. Eng. Mater. Technol. 98, 6068.
ASTM E45, 1997. Standard test methods for determining the inclusion content of
steel. In American Society for Testing and Materials. 3.01. ASTM, Philadelphia,
Pennsylvania, pp. 163176.
Bai, Y., Dodd, B., 1992. Adiabatic Shear Localization: Occurrences, Theories and Appli-
cations. Pregamon Press, New York, pp. 470.
Bandstra, J.P., Koss, D.A., Geltmacher, A., Matic, P., Everett, R.K., 2004. Modeling void
coalescence during ductile fracture of a steel. Mater. Sci. Eng. A 366A, 269
281.
Barrett, R., 1997. Getting a grip on fasteners. Metal Bulletin Monthly. February, pp.
3843.
Boyer, J.-C., Vidal-Sall, E., Staub, C., 2002. A shear stress dependent ductile damage
model. J. Eng. Mater. Technol. 121, 8793.
Broek, D., 1973. The role of inclusions in ductile fracture and fracture toughness. Eng.
Fract. Mech. 5, 5556.
Chen, Z.H., Tang, C.Y., Lee, T.C., Chan, L.C., 2002. Numerical simulation of ne-
blanking process using a mixed nite element method. Int. J. Mech. Sci. 44, 1309
1333.
Chen, Z.H., Chan, L.C., Lee, T.C., Tang, C.Y., 2003. An investigation on the formation
and propagation of shear band in ne-blanking process. J. Eng. Mater. Technol.
138, 610614.
Chitkara, N.R., Bhutta, M.A., 2001. Shape heading of splines and solid spur gear forms:
an analysis and some experiments. Int. J. Mech. Sci. 43 (4), 10731106.
Cockroft, M.G., Latham, D.J., 1968. Ductile and workability of metals. J. Inst. Met. 96,
3339.
Cowie, J.G., Azrin, M., Olson, G.B., 1989. Microvoid formation during shear deforma-
tion of ultrahigh strength steels. Metall. Trans. A: Phys. Metall. Mater. Sci. 20A
(1), 143153.
Cox, T.B., Low, J.R., 1974. An investigation of the plastic fracture of AISI 4340 and 18
nickel-200 grade maraging steels. Metall. Trans. A: Phys. Metall. Mater. Sci. 5 (6),
14571470.
Daridon, L., Oussouaddi, O., Ahzi, S., 2004. Inuence of the material constitutive
models on the adiabatic shear band spacing: MTS, power law and JohnsonCook
models. Int. J. Sol. Struct. 41 (1112), 31093124.
Dodd, B., Atkins, A.G., 1983. Flow localization in shear deformation of void-
containing and void-free solids. Acta Metall. 31 (1), 915.
Dodd, B., Bai, Y., 1987. Ductile Fracture and Ductility with Applications to Metalwork-
ing. Academic Press Inc, London.
Dodd, B., Hartley, P., 1991. Workability and CAD/CAM. J. Mater. Proc. Technol. 26 (1),
3552.
Garrison Jr., W.M., Moody, N.R., 1987. Ductile fracture. J. Phys. Chem. Sol. 48 (11),
10351074.
Gurland, J., 1972. Observations on the fracture of cementite particles in a
spheroidized 1.05% C steel deformed at room temperature. Acta Metall. 20,
735741.
Hahn, G.T., Roseneld, A.R., 1975. Metallurgical factors affecting fracture toughness
of aluminum alloys. Metall. Trans. 6A, 653670.
Hambli, R., 2001. Finite element simulation of ne blanking processes using a
pressure-dependent damage model. J. Mater. Proc. Technol. 116 (23, 24),
252264.
Horstemeyer, M.F., Gokhale, A.M., 1999. A voidcrack nucleation model for ductile
metals. Int. J. Sol. Struct. 36 (33), 50295055.
Kawashima, T., 1992. Cold heading for special products. Wire Ind. (UK) 60 (714),
340345.
Klepaczko, J.R., Lipinski, P., Molinari, A., 1988. An analysis of the thermoplastic catas-
trophic shear in some metals. DGM Informationsgesellschaft MBH. Impact Load.
Dyn. Behav. Mater. 2, 695704.
Lee, P.W., Kuhn, H.A., 1973. Fracture in cold upset forginga criterion and model.
Metall. Trans. 4, 969973.
Lee, W.S., Lin, C.F., Liu, C.Y., Tzeng, F.T., 2004. Impact properties of 304L stainless steel
GTAW joints evaluated by high strain rate of compression tests. J. Nucl. Mater.
335 (3), 335344.
Masheshwari, M.D., Dutta, B., Mukherjee, T., 1978. Quality requirements for cold
heading grades of steel. Tool Alloy Steels, 247251.
Matsunaga, T., Shiwaku, K., 1980. Manufacturing of cold heading quality wire rods
and wires. SEAISI Quart. 9 (1), 4555.
Meyers, M.A., Nesterenko, V.F., La Salvia, J.C., Xue, Q., 2001. Shear localization
in dynamic deformation of materials: microstructural evolution and self-
organization. Mater. Sci. Eng. A 317A, 204225.
 A. Sabih, J.A. Nemes / Journal of Materials Processing Technology 209 (2009) 42924311
Nemat-Nasser, S., Guo, W.G., Kihl, D.P., 2001. Thermomechanical response of AL-6XN
stainless steel over a wide range of strain rates and temperatures. J. Mech. Phys.
Sol. 49 (8), 18231846.
Nickoletopoulos, N., 2000. Physical and Numerical Modeling of Fracture During
Upsetting for Cold Heading Operations. Doctoral Dissertation. McGill University
at Montreal.
Nickoletopoulos, N., Hone, M., Wanjara, P., Verreman, Y., Nemes, J.A., Jonase, J.J., Yue,
S., 2001. Development of a fracture criterion for cold heading. Wire J. Int. 34 (9),
9098.
Nieh, T.G., Nix, W.D., 1980. A comparison of the dimple spacing on intergranular
creep fracture surfaces with the slip band spacing for copper. Scr. Metall. 14 (3),
365368.
Okamoto, T., Fukuda, T., Hagita, H., 1973. Material fracture in cold forgingsystematic
classication of working methods and types of cracking in cold forging. Sumit-
omo Search 9, 216227.
Pardoen, T., Delannay, F., 1998. Assessment of void growth models from porosity
measurement in cold-drawn copper bars. Metall. Mater. Trans. A: Phys. Metall.
Mater. Sci. 29A, 18951909.
Park, I.G., Thompson, A.W., 1988. Ductile fracture in spheroidized 1520 steel. Acta
Metall. 36 (7), 16531664.
Petersen, S.B., Rodrigues, J.M.C., Martins, P.A.F., Lopes, A.B., Grcio, J.J., 1997. Injec-
tion forging of tubular materials: workability analysis. J. Mater. Proc. Technol. 65
(13), 8893.
Pierson, G., 1986. Improvements and developments in the cold heading eld. Wire
J. Int. 19, 5152.
Pirondi, A., Bonora, N., 2003. Modeling ductile damage under fully reversed cycling.
Comp. Mater. Sci. 26, 129141.
Puttick, K.E., 1959. Ductile fracture in metals. Philos. Mag. 4, 964969.
Qiang, L., Bassim, M.N., 2003. Effects of strain and strain-rate on the formation of the
shear band in metals. J. Phys. IV France 110, 8791.
Reitz, W., 2005. Failure analysis of brass bolt from mausoleumcase study. JFAP 5
(4), 2025.
Rogers, H.C., 1959. The mechanism of crack propagation in ductile metals. Acta Met-
all. 7 (11), 750752.
Rogers, H.C., 1960. The tensile fracture of ductile metals. Trans. Metall. Soc., AIME
Trans. 218, 498506.
Rogers, H.C., 1983. Adiabatic shearinggeneral nature and material aspects. In:
Mescall, J., Weiss, V. (Eds.), Material Behavior Under High Stress and Ultrahigh
Loading Rates. Sagamore Army, Materials Research Conference, Lake Placid. New
York, pp. 101118.
Sabih, A., Wanjara, P., Nemes, J.A., 2005. Characterization of internal voids and cracks
in cold heading of dual phase steel. ISIJ Int. 45 (8), 11791186.
4311
Sabih, A., El-Wazri, A.M., Nemes, J.A., Yue, S., 2006a. A workability criterion for the
transformed adiabatic shear band phenomena during cold heading of 1038 steel.
JFAP 6 (4), 97105.
Sabih, A., Nemes, J.A., Nickoletopoulos, N., Hone, M., 2006b. A workability criterion
for internal ductile fracture in the steel cold heading process. Wire J. Int. 39 (3),
155159.
Sarruf, Y., Cao, B., Jonas, J.J., Nickoletopoulos, N., 1999. Criteria and test for cold
headability. Wire J. Int. 32 (2), 98105.
Siruguet, K., Leblond, J.B., 2004a. Effect of void locking by inclusions upon the plastic
behavior of porous ductile solids. I. Theoretical modeling and numerical study
of void growth. Int. J. Plast. 20 (2), 225254.
Siruguet, K., Leblond, J.B., 2004b. Effect of void locking by inclusions upon the plastic
behavior of porous ductile solids. Part II. Theoretical modeling and numerical
study of void coalescence. Int. J. Plast. 20 (2), 255268.
Tanguy, B., Bouchet, C., Bugat, S., Besson, J., 2006. Local approach to fracture based
prediction of the T56J and TKIc,100 shifts due to irradiation for an A508 pressure
vessel steel. Eng. Fract. Mech. 73 (2), 191206.
Teng, X., Wierzbicki, T., Couque, H., 2007. On the transition from adiabatic shear
banding to fracture. Mech. Mater. 39 (2), 107125.
Thomson, P.F., 1990. A Ductile Fracture of Metals. Pergamon Press, New York, pp.
3039.
Tvergaard, V., 1981. Inuence of voids on shear band instabilities under plane strain
conditions. Int. J. Fract. 17, 381388.
Valiente, A., Ruiz, J., Elices, M., 2005. A probabilistic model for the pearlite-induced
cleavage of a plain carbon structural steel. Eng. Fract. Mech. 72 (5), 709728.
Venugopal, S.L., Mannan, Y.V., Prasad, R.K., 1997. Instability map for cold and warm
working of as-cast 304 stainless steel. J. Mater. Proc. Technol. 65 (13), 107115.
Vora, H., Polonis, D.H., 1976. Inuence of microstructural features on the dynamic
fracture behavior of some iron-base materials. Mater. Sci. Eng. 22, 5160.
Wei, Q., Jiao, T., Ramesh, K.T., Ma, E., Kecskes, L.J., Magness, L., Dowding, R.,
Kazykhanov, V.U., Valiev, R.Z., 2006. Mechanical behavior and dynamic failure
of high-strength ultrane grained tungsten under uniaxial compression. Acta
Mater. 54 (1), 7787.
Wright, T.W., 2002. The Physics and Mathematics of Adiabatic Shear Bands. Cam-
bridge Monographs on Mechanics. Cambridge University Press, Cambridge, pp.
133.
Yoo, S.J., Lee, D.L., Chung, D.T., 1997. Prediction of cold heading quality with high
strain rate compression tester. Wire J. Int. 30 (9), 8489.
Zhang, L., Thomas, B., 2002. Evaluation and control of steel cleanlinessreview. In:
Proceedings of the 85th Steelmaking Conference, ISS-AIME, Warrendale, PA, pp.
431452.
Zukas, J., 1990. High Velocity Impact Dynamics. John Wiley and Son, New York.